Hull resistance prediction with nDAI

Within a hull optimization process aimed at reducing resistance over a range of operating speeds, a reduced-order model (ROM) has been trained with nDAI. In the first offline phase, the CAD design variables are redefined through the Parametric Model Embedding (PME) technique. On this basis, a Design of Experiments (DOE) is generated with modeFRONTIER and used to drive geometry generation in NX and CFD simulations in STAR-CCM+. In the second offline phase, the resulting hull geometries are then aligned and analyzed to extract a small set of key parameters that capture the main shape variations. The ROM model is then built to support the designer in visualizing pressure, shear stresses, and volume of fluid distributions on the hull as functions of the key parameters.